Page 145 - Read Online

Park et al. J Cancer Metastasis Treat 2019;5:17 I http://dx.doi.org/10.20517/2394-4722.2018.84 Page 5 of 12

with FITC-conjugated antibody to ICAM-1 in the presence of Fc Block for 30 min at 4 °C. Cells were then
washed 3× with PBS and analyzed by FACS Caliber (BD). Data analysis was performed using Flowjo V10
software (Ashland).

Monocyte adhesion assay
Isolated T cells were incubated with Cell Tracker (Thermo Fisher Scientific) for 30 min at 37 °C and then
5
10 cells per well were incubated on MDA-MB-231LN monolayers in HBSS supplemented with 2 mmol/L
calcium and magnesium for 1 h. After removing nonadherent cells, the number of Cell Tracker-labeled cells
bound to the MDA-MB-231LN cells was quantified under fluorescent microscopy.

Immunohistochemistry
Tissue samples were fixed in 4% paraformaldehyde, embedded in paraffin, and then 4-µm sections were
prepared. Sections were de-waxed and a steamer pre-treatment in Tris/EDTA buffer (DAKO) was performed.
Endogenous peroxidase activity was quenched by incubation in 3% hydrogen peroxide in PBS. For blocking
steps, avidin (Sigma-Aldrich), biotin (Sigma-Aldrich) in PBS, and a super block (IDlabs Biotechnology) were
used. Rabbit monoclonal AF1q antibody and mouse monoclonal ICAM-1 antibody in a 1:200 dilution were
incubated at 4 °C overnight. IHC detection was performed with the IDetect Super Stain System HRP (IDlabs
Biotechnology). Specific signals were amplified using 3-amino-9-ethylcarbazole (IDlabs Biotechnology)
under visual control, followed by counterstaining with hematoxylin.

Statistical analysis
Quantitative data are presented as means ± SD and were subjected to analysis of variance followed by a t test
with the use of Prism V.7 software (GraphPad Software). A P value of < 0.05 was considered statistically
significant.

RESULTS
Next-generation sequencing and transcriptome data analysis
We observed AF1q, a metastasis enhancer, was highly expressed in metastatic cells, MDA-MB-231LN
(invasive subline from MDA-MB-231), than in the primary tumor cells (MDA-MB-231) [Figure 1A]. To
validate the results of RNA-seq, we performed RT-qPCR analysis using randomly selected genes [Figure 1B].
Both MDA-MB-231 and MDA-MB-231LN cell lines were subjected to RNA-Seq analysis with three biological
replicates for each cell line. The 75 bp-long single-end sequencing reads were mapped to the human genomic
reference using STAR aligner. More than 27.7 million sequencing reads were mapped to genome on each
replicate (averaging 35.3 million mapped reads) [Supplementary Table 1]. Out of the 58,289 annotated genes,
18,042 (31.0%) by Ensemble were expressed with more than 10 read counts on at least one of our replicates
[Supplementary Table 2]. Only uniquely mapped reads were used further for searching differentially
expressed genes using Cuffdiff. Total 1,485 genes were selected as significantly differentially expressed with
the criteria explained above. Among those, 762 genes showed increased value of RPKM in MDA-MB-231LN
cell line compared to MDA-MB-231 and 723 genes showed the opposite. Row-scaled RPKM value of each
replicate represents that the differential expression of genes is consistent throughout replicates of each cell
line [Figure 1C].

With those differentially expressed genes, pathway analysis was performed using Metacore. By selecting
top 40 network objects resulted from the pathway analysis [Table 1], direct interaction network building
algorithm was applied. Interestingly, pathway process showed that most genes were directly linked with
ICAM-1 [Figure 2A]. In addition, we found out that ICAM-1 transcript level was significantly decreased in
MDA-MB-231LN cell line compared to MDA-MB-231 confirming that ICAM-1 expression is attenuated in
metastatic cell line [Figure 2B].

To explore the relationship of AF1q with ICAM-1 in breast cancer, we used 1,222 RNA-seq data of 1,092
breast cancer cases from TCGA database. Using FPKM values from all 1,222 samples, the expression level of